改进的遗传算法在催化裂解集总动力学参数估值中的应用  

Application of an improved genetic algorithm in estimating lumped kinetic parameters for catalytic cracking reaction

作　　者：许闽[1] 朱吉钦[1] 刘辉[1] 贺振富[2] 田辉平[2] 

机构地区：[1]北京化工大学化工资源有效利用国家重点实验室35信箱,北京100029 [2]中国石油化工股份有限公司石油化工科学研究院,北京100083

出　　处：《计算机与应用化学》2008年第10期1225-1228,共4页Computers and Applied Chemistry

基　　金：国家重点基础研究发展规划项目(2004CB719505)

摘　　要：以基本遗传算法为基础,引入多重退火交叉,并以多种变异模式竞争取代单一的变异策略,以随着遗传代数和个体适应度动态调节的交叉和变异概率代替固定不变的交叉和变异概率,提出了一种改进的遗传算法;用于研究包含复杂组分、同时进行多种反应的催化裂解反应集总动力学,并与传统的算法比较,证明改进的遗传算法可迅速、准确获得有物理意义的动力学参数优化值,是研究复杂反应动力学的有效数值工具。An improved genetic algorithm （GA） was applied to the estimation of 5 lumped model parameters appearing in the rate equations of catalytic cracking reaction. The GA was developed by a hybrid use of such strategies as multi-annealing crossover, competitive mutation, and crossover probability and mutation probabilities varying with generations and fitness. It is demonstrated that the GA algorithm yields, with a higher estimating precision and better convergence than the conventional GA algorithms, the optimal values of the rate parameters for the complex reaction system, thereby constituting a useful numerical tool for studying complex reaction kinetics.

关 键 词：遗传算法 催化裂解 集总动力学 

分 类 号：TQ203.3[化学工程—有机化工]